弥漫大B细胞淋巴瘤患者1年内死亡相关风险因素及预测模型的构建  

Risk factors related to the death of patients with diffuse large B-cell lymphoma within one year and the construction of prediction model

作　　者：李娜娜 丁彬彬 董白 李子坚[2] LI Nana;DING Binbin;DONG Bai;LI Zijian(The First School of Clinical Medicine,Lanzhou University,Gansu Lanzhou 730000,China;Department of Hematology,the First Hospital of Lanzhou University,Gansu Lanzhou 730000,China)

机构地区：[1]兰州大学第一临床医学院,甘肃兰州730000 [2]兰州大学第一医院血液科,甘肃兰州730000

出　　处：《现代肿瘤医学》2025年第2期263-268,共6页Journal of Modern Oncology

基　　金：甘肃省兰州市科技计划(编号:2020-ZD-68)。

摘　　要：目的:基于最小绝对收缩与选择算法(minimum absolute shrinkage and selection algorithm,LASSO)回归探讨弥漫大B细胞淋巴瘤(diffuse large B-cell lymphoma,DLBCL)患者1年内死亡相关危险因素并构建预测模型。方法:收集2013年01月至2024年06月就诊于兰州大学第一医院确诊弥漫大B细胞淋巴瘤患者临床资料,按1年内是否死亡分为死亡组和生存组,比较患者的临床资料,分析其死亡的相关危险因素,基于LASSO回归分析筛选变量,以特征性变量作为自变量代入多因素Logistic回归模型获得弥漫大B细胞淋巴瘤患者1年内死亡的独立预测因素并构建预测模型,绘制ROC曲线、校准曲线、临床决策曲线(decision curve analysis,DCA)来评估模型的预测效果。结果:LASSO回归筛选出最具泛化能力的4个特征性变量:Ki67、ECOG评分、完成6个周期化疗、IPI评分,其对应的LASSO回归系数分别为0.506、0.483、-0.855及2.595。多因素Logistic回归分析调整和校正后,Ki67、ECOG评分、6个周期化疗、IPI评分为弥漫大B细胞淋巴瘤患者1年内发生死亡的独立预测因素(P<0.05),构建列线图,列线图模型的ROC曲线显示曲线下面积(area under the curve,AUC)为0.944(95%CI:0.906~0.983),校准曲线显示Hosmer-Lemeshow拟合优度检验(χ2=9.1856,df=7,P=0.2396)表明模型有良好的拟合度,DCA曲线显示该模型具有明显正向净收益。结论:Ki67、ECOG评分、完成6个周期化疗、IPI评分为弥漫大B细胞淋巴瘤患者1年内发生死亡的独立预测因素,以这些变量构建的临床预测模型能较好预测弥漫大B细胞淋巴瘤的预后及识别死亡风险人群。Objective:To investigate the risk factors related to death with 1 year in diffuse large B-cell lymphoma(DLBCL)and construct a prediction model based on the minimum absolute shrinkage and selection algorithm(LASSO)regression.Methods:Collecting the clinical data of patients with confirmed diffuse large B-cell lymphoma treated in the First Hospital of Lanzhou University from January 2013 to June 2024.The patients were divided into death and survival groups according to whether death within 1 year.Comparing the clinical data of the patients and analyzing the related risk factors for their death.Screating variables based on minimum absolute shrinkage and selection algorithm,to obtain independent predictors of death within 1 year in diffuse large B-cell lymphoma patients with a characteristic variable as an independent variable,and to construct a prediction model.The ROC curve,calibration curve,and clinical decision curve analysis(DCA)were drawn to evaluate the prediction effect of the model.Results:LASSO regression selected the four characteristic variables with the most generalization ability:Ki67,ECOG score,6 cycles of chemotherapy completed,and IPI score.The corresponding LASSO regression coefficients were 0.506,0.483,-0.855 and 2.595,respectively.After adjustment and correction for the multivariate Logistic regression analysis,Ki67,ECOG score,6 cycles of chemotherapy,IPI score were independent predictors of death within 1 year in patients with diffuse large B-cell lymphoma(P<0.05).Building the nomogram,the ROC curve of the nomogram model showed that the area under the ROC curve(AUC)was 0.944(95%CI:0.906 to 0.983).The calibration curve showed the Hosmer-Lemeshow goodness of fit test(χ^(2)=9.1856,df=7,P=0.2396).The DCA curve showed that the model had a clear positive net gain.Conclusion:Ki67,ECOG score,6 cycles of chemotherapy completed,and IPI score are independent predictors of death within 1 year in patients with diffuse large B-cell lymphoma.The clinical prediction model constructed with these variables can bet

关 键 词：弥漫大B细胞淋巴瘤 死亡风险 危险因素 预测模型 

分 类 号：R733[医药卫生—肿瘤]